INDEX

Explanations

the word "be" at high activations, indicating a focus on statements of existence or identification

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "be" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 illuminate

-0.69

anwhile

-0.69

imize

-0.67

rones

-0.67

ipeg

-0.66

rouse

-0.66

culosis

-0.65

ãĥ¼ãĥ«

-0.64

 borrow

-0.64

ppings

-0.64

POSITIVE LOGITS

 able

0.99

leeve

0.87

 regarded

0.83

hemoth

0.82

heading

0.81

getting

0.81

 considered

0.78

fits

0.77

 influenced

0.74

fit

0.74

Activations Density 0.105%

the word "be" at high activations, indicating a focus on statements of existence or identification

instances of the verb "be" in various contexts

No Comments

No Known Activations